A Coordination and Discovery Service for QoS-enabled Data-Centric Publish/Subscribe in Wide Area Networks

A special class of Internet of Things called the Industrial Internet of Things (IIoT) operates in a large, distributed and dynamic environment comprising sensors all the way to large server clusters. A key requirement for IIoT is a scalable messaging service that supports multiple quality of service (QoS) properties, such as timeliness and resilience. Although existing pub/sub standards, such as the Object Management Group (OMG)’s Data Distribution Service (DDS) for data-centric pub/sub, support a range of QoS properties and dynamic discovery of peers, they are effective only in local area networks (LANs). Moreover, they lack an effective coordination and discovery service needed by brokers that can bridge multiple different LANs. To address these limitations, this paper presents PubSubCoord, which is a cloud-enabled coordination and discovery service for QoSenabled data-centric pub/sub for wide area network (WAN) operations. PuSubCoord realizes a WAN-scale, low-latency data dissemination architecture by (a) balancing the load using elastic cloud resources, (b) clustering brokers by topics for affinity, and (c) minimizing the number of data delivery hops in the pub/sub overlay. PubSubCoord’s coordination mechanism uses ZooKeeper to support dynamic discovery of brokers and pub/sub endpoints located in isolated networks. Empirical results evaluating the performance of PubSubCoord are presented for (1) scalability of data dissemination and coordination, and (2) deadline-aware overlays employing configurable QoS to provide low-latency data delivery for topics demanding strict service requirements. Keywords-Data Distribution Service, Publish/Subscribe, Middleware, Discovery, Coordination, Cloud Computing